Shedding light on the copper-catalysed diboron(4) reduction of nitrous oxide

Abstract

Catalytic deoxygenation of the potent greenhouse gas and ozone-depleting agent N₂O mediated by NHC-ligated copper(I) boryl complexes has been examined under a variety of reaction conditions, including different diboron(4) reducing agents, and benchmarked against a rhodium(I) system. While unstable and decomposing rapidly in light, complexes of SIMes and IMes deliver the highest catalytic activity in combination with B₂pin₂ as the reducing agent when performed in the dark using THF as the solvent, achieving ∼2000 TONs over 20 h at room temperature under 1 bar gauge of N₂O pressure. DFT-based computational analysis corroborates a mechanism involving reaction of the copper(I) boryl with N₂O by O-atom insertion into the Cu–B bond (via initial κ_N-coordinaton) followed by sigma-bond metathesis between the resulting boroxide derivative and diboron(4) reducing agent, with the relative barriers nuanced by the nature of the supporting NHC ligand and solvent employed.

---

Introduction

Nitrous oxide (laughing gas, N₂O) is the third most abundant greenhouse gas, with a global warming potential 273 times greater than carbon dioxide, and the dominant ozone depleting substance emitted in the 21st century.¹ Exponentially increasing anthropogenic emissions make it imperative that methods for the remediation and/or repurposing of nitrous oxide are developed, but controlled activation of this atmospheric pollutant has proven to be a challenging problem.² For homogeneous late transition metal complexes, insertion into covalent M–X bonds (X = H, C, B) has emerged as a promising strategy,³ as exemplified by the use of ruthenium and rhodium hydride complexes as (pre-)catalysts for the hydrogenation of N₂O.⁴ Building upon stoichiometric work by Braun using [(Et₃P)₃Rh(Bpin)] (1, Scheme 1A),⁵ and recognising the propensity of boron to form very strong bonds with oxygen (D_e = 809 kJ mol⁻¹),⁶ we have recently set about investigating the use of late transition metal boryl complexes as homogeneous catalysts for the deoxygenation of N₂O.

Scheme 1 (A) Stoichiometric and (B) catalytic deoxygenation of N₂O mediated by metal boryl complexes using B₂pin₂ as the reductant (pin = pinacolato).

In preceding work, carried out in collaboration with Liptrot and inspired by the homogeneous process developed by Sadighi for the reduction of isoelectronic CO₂,⁷ we showed that NHC-ligated copper(I) tert-butoxide complexes are effective pre-catalysts for the deoxygenation of N₂O to N₂ using the diboron(4) compound B₂pin₂ (pin = pinacolato) as the reductant in benzene.⁸ Robust catalytic performance was noted for [(SIPr)Cu(OtBu)] 2* and the proposed mechanism, involving reaction of the corresponding copper(I) boryl [(SIPr)Cu(Bpin)] 2 with N₂O by O-atom insertion into the Cu–B bond to liberate N₂ followed by rate determining sigma-bond metathesis between [(SIPr)Cu(OBpin)] 2′ and B₂pin₂, was established for this system experimentally (Scheme 1B). Guided by protocols reported in the literature,^7,9 all reactions were performed in the dark to mitigate against the inferred sensitivity of the catalytically active copper(I) boryls to light. Catalyst decomposition was however noted during initial catalyst screening, performed within high-pressure J. Young value NMR tubes shielded from light (with foil when not in the spectrometer), particularly in the case of the Mes-substituted NHC systems examined. Productive photoactive copper(I) systems are typically tetracoordinated.¹⁰

Results and discussion

With a view of optimising the experimental conditions, the copper(I)-catalysed deoxygenation of N₂O was re-examined for a homologous series of tert-butoxide precatalysts, where the supporting NHC ligand is SIPr (2*), IPr (3*), SIMes (4*), or IMes (5*),¹¹ using commercially available B₂pin₂, B₂neop₂ (neop = neopentyl glycolato), or B₂cat₂ (cat = catecholato) as the diboron(4) reducing agent and THF or toluene as the solvent (Table 1A and B). Catalytic activity was assessed using 5 mol% pre-catalyst (5 mM) at room temperature under 1/3 bar gauge N₂O (1/3 bar absolute N₂O, balance argon) in the dark. Experiments were performed in parallel using a bespoke multiwell stainless-steel pressure reactor (Table 1C), with each individual reaction conducted within an amberised glass vessel that was charged with the pre-catalyst, diboron(4) and solvent in an argon filled glovebox in the dark to ensure the most rigorous exclusion of light possible. For comparison, the catalytic activity of unligated copper tert-butoxide 6* and [(Et₃P)₃Rh(OPh)] 1*¹² were determined under selected reaction conditions. Control experiments verified that, the diboron(4) reagents do not react with N₂O under the conditions studied.

Table 1 Catalyst screening for the diboron(4) reduction of N₂O^a

Entry	Conditions				Conversion/%
	Diboron(4)	Solvent	pN2O/bar	Time/min	1*	2*	3*	4*	5*	6*
a Conditions: 5 μmol precatalyst and 100 μmol of diboron(4) in 1 mL of solvent. Individual samples prepared in the dark within amberised glass vessels and parallel reactions run at room temperature inside a stainless-steel pressure reactor (1 atm argon, pressured to 1/3 bar gauge N2O). Conversion determined by ^11B NMR spectroscopy and averaged over duplicate runs.
1	B2pin2	THF	3	120	6	34	87	100	100	22
2	B2pin2	THF	3	10				99	92
3	B2pin2	THF	1	120		33	90	100	100
4	B2pin2	THF	1	10				96	93
5	B2pin2	Toluene	3	120		81	99	100	100	21
6	B2pin2	Toluene	3	10				58	84
7	B2pin2	Toluene	1	120		81	100	100	100
8	B2pin2	Toluene	1	10				42	63
9	B2neop2	THF	3	120	35	77	83	52	63	12
10	B2neop2	Toluene	3	120		87	74	41	36	16
11	B2cat2	THF	3	120	88	25	29	<5	<5	0
12	B2cat2	Toluene	3	120		19	13	20	13	<5

---

This expanded catalyst screening reaffirms that NHC-ligated copper(I) tert-butoxide complexes are effective pre-catalysts for the deoxygenation of N₂O, using diboron(4) compounds as the reducing agent that afford bis(boryl)oxides as the boron-containing byproducts of the reaction (alongside boryl tert-butoxide, see SI). Using our refined protocol, the Mes-substituted NHC pre-catalysts 4* and 5* are found to the most active N₂O deoxygenation catalysts when using B₂pin₂ as the reducing agent and p_N2O = 1 or 3 bar gauge, with complete consumption of diboron(4) observed in both THF and toluene within 2 h. Repeating these reactions under reduced turnover conditions revealed a N₂O pressure dependency in toluene, and enabled a lower activity limit of TOF_avg > 40 h⁻¹ to be established. In combination with the less bulky alkyl diboron(4) B₂neop₂, these pre-catalysts do, however, show reduced activity and are outperformed by the Dipp-substituted NHC-precatalysts 2* and 3* (Table 1, entries 9 and 10), which are otherwise notable for enhanced activity in toluene with B₂pin₂. The copper(I) complexes are found to be least effective in catalysis when B₂cat₂ was used as the reducing agent and this outcome is attributed to detrimental reactions of the associated boryl derivatives initiated by Lewis acids (e.g. tBuO(Bcat), O(Bcat)₂).¹³ In this context, it is interesting to note that the rhodium(I) pre-catalyst 1* performs best in combination with B₂cat₂ and the associated deoxygenation activity (TOF_avg ∼ 9 h⁻¹) can be reproduced for pincer analogues of the form [Rh(pincer)X] (pincer = 2,6-(iPr₂PCH₂)₂C₅H₃N, X = OPh; Xantphos-iPr, X = Bpin; see SI).¹⁴

To help understand the influence of light in catalysis, we have systematically studied the stability of the copper(I) boryl complexes [(NHC)Cu(Bpin)] 2–5 (20 mM), generated in situ from reaction of the pre-catalysts [(NHC)Cu(OtBu)] (2*–5*) with B₂pin₂ in THF and toluene, with and without precautions for the exclusion of light (Table 2). In our hands, the Dipp-substituted NHC boryls 2 and 3 are stable in the dark and undergo only slow decomposition when exposed to light, with [(SIPr)Cu(Bpin)] practically light stable when prepared in THF (Table 2, entry 3). While the less bulky Mes-substituted NHC boryls 4 and 5 can be unambiguously identified in situ by ¹¹B NMR spectroscopy (ca. δ_11B 42) when prepared in the dark, both are unstable and decomposed extremely rapidly into an intractable mixture of species in the light. Overall, the stability of the boryl complexes decreases in the order 2 > 3 ≫ 4 > 5 and THF > toluene and these trends vindicate re-examining the catalytic activity of the copper pre-catalysts with more rigorous exclusion of light in different solvents. Running catalytic reactions under more dilute conditions, as we have here, may also lead to some discrepancies, as this would help suppress bimolecular catalyst decomposition pathways. Such processes will be more apparent at the relatively high copper concentrations used in this stability study to facilitate analysis by multinuclear NMR spectroscopy.

Table 2 Stability of boryl derivatives 2–5 generated in situ from the reaction between [(NHC)Cu(OtBu)] 2*–5* and B₂pin₂ at room temperature^a

Entry	Conditions		Stability [(NHC)Cu(Bpin)]
	Solvent	Light	2 (SIPr)	3 (IPr)	4 (SIMes)	5 (IMes)
a Conditions: 10 μmol 2–5 and 11 μmol of B2pin2 in 0.5 mL of either d8-THF or d8-toluene. Samples prepared in the dark within either an amberised or clear glass J. Young valve NMR tube. Ambient = interior laboratory lighting, direct sunlight = placed adjacent to a window on a sunny day.
1	THF	Dark	Stable	Stable	t ∼ 30 min	t < 10 min
2	THF	Ambient	Stable	t > 24 h	t < 10 min	Not observed
3	THF	Direct sunlight	Stable	t1/2 ∼ 10 h
4	Toluene	Dark	Stable	Stable	t ∼ 30 min	t < 5 min
5	Toluene	Ambient	Stable	t > 24 h	t < 5 min	Not observed
6	Toluene	Direct sunlight	t > 24 h	t1/2 ∼ 3 h

---

Given the large differences in light sensitivity evident for 2–5, we turned to computational methods to interrogate the mechanistic subtleties in N₂O deoxygenation catalysis, focusing on the most effective diboron(4) reducing agent B₂pin₂ and selecting DFT calculations at the B3LYP-D3(BJ)/def2-TZVP//BP86-D3(BJ)/def2-SVP level of theory corrected for benzene, toluene or THF solvent (SMD).¹⁵ Leveraging our preceding work in benzene as a robust experimental benchmark,⁸ the reaction profile for SIPr-ligated 2 was analysed in the first instance. Particularly to assess the relative energetics of O-atom transfer pathways, involving concerted insertion into the Cu–B bond or, informed by related computational work,^16,17 addition of N₂O across the Cu–B bond (Fig. 1A). The former is associated with a prohibitively large barrier of ΔG^‡_298K = 31.9 kcal mol⁻¹ (2/TS_O) while two stepwise pathways for the latter, where the terminal O atom of N₂O approaches either the Cu (ΔG^‡_298K = 18.9 kcal mol⁻¹, 2/TS_ON2) or B (ΔG^‡_298K = 16.8 kcal mol⁻¹, 2/TS_N2O) centres, could be identified. The most favourable variant can be interpreted as a nucleophilic attack of the boryl at the pendant O atom of κ_N-coordinated N₂O, as can be visualised by EDA-NOCV analysis of associated transition state 2/TS_N2O (Fig. 1B).¹⁸ Binding of N₂O along this pathway is substantiated by QTAIM analysis (Fig. S171) and, while endergonic in this case, there is experimental precedent for intact κ_N-coordinaton of N₂O to copper(I).¹⁹

Fig. 1 (A) Computed reaction profile for deoxygenation of N₂O catalysed by 2 using B₂pin₂ as the reducing agent, with (B) EDA-NOCV analysis of the lowest energy O-atom transfer transition state. Calculations at the B3LYP-D3(BJ)/def2-TZVP//BP86-D3(BJ)/def2-SVP level of theory corrected for benzene solvent (SMD).¹⁵

The computed thermodynamics indicate that subsequent liberation of N₂ and formation of the resulting copper(I)-boroxide 2′ is highly exergonic, with ΔG_298K = −117.8 kcal mol⁻¹ relative to 2 and N₂O. The onward sigma-bond metathesis step has previously been studied computationally by Lin and Mayer in the context of CO₂ reduction,¹⁶ and the overall activation barrier (ΔG^‡_298K = 21.3 kcal mol⁻¹, 2′/TS_SBM) and thermodynamics (ΔG_298K = −5.4 kcal mol⁻¹, relative to 2 + B₂pin₂; cf. −123.2 kcal mol⁻¹ for the overall transformation) at our chosen level of theory are in good agreement. Critically, the calculated activation barrier for this step is significantly larger than that of the O-atom transfer step (ΔΔG^‡_298K = 4.5 kcal mol⁻¹), consistent with the experimental observation that this is rate determining (viz. no N₂O pressure dependence was observed during the screening).

Informed by this analysis, the activation barriers for the O-atom transfer and sigma-bond metathesis steps were computed for the complete homologous series of copper(I)-boryl catalysts 2–5, as pertinent to the catalytic reactions performed in toluene and THF solvent (Table 3). O-atom transfer via TS_N2O is most favourable for all systems (Table S6), with the calculated barriers for the Dipp-substituted NHC catalysts 2 and 3 marginally lower than the Mes-substituted analogues 4 and 5 (ΔΔG^‡_298K ca. –2 kcal mol⁻¹, Table 3). Consistent with the relatively low activity observed for 2* (Table 1, entries 1 and 5), rate determining sigma-bond metathesis between SIPr-ligated 2′ and B₂pin₂ invokes a barrier ca. 3 kcal mol⁻¹ higher than for 3′–5′. Moreover, the significantly higher activity observed for 2* in toluene vs. THF is reproduced computationally. Higher barriers for sigma-bond metathesis are calculated in THF vs. toluene across the board due to electrostatic stabilisation of the ground state in THF. Less pronounced solvent effects are found for the O-atom transfer barriers and has significant mechanistic implications for the Mes-substituted NHC catalysts 4 and 5, for which the barriers for O-atom transfer and sigma bond metathesis in toluene are calculated to be within 0.5 kcal mol⁻¹, cf. the barrier for sigma-bond metathesis being >1 kcal mol⁻¹ than O-atom transfer for 2 and 3 in toluene and 2–5 in THF. On this basis, the reduced catalytic activity and N₂O pressure dependency in toluene vs. THF observed experimentally for 4 and 5 is attributed to the deoxygenation of N₂O becoming a pseudo third order reaction in toluene.

Table 3 Calculated activation barriers ΔG^‡_298K/kcal mol⁻¹ for the deoxygenation of N₂O catalysed by 2–5 using B₂pin₂ as the reducing agent

Catalyst	NHC	THF			Toluene
		OAT	SBM	Δ	OAT	SBM	Δ
2	SIPr	16.0	23.0	7.0	16.7	21.2	4.5
3	IPr	16.4	19.9	3.5	17.0	18.4	1.4
4	SIMes	17.4	20.0	2.6	17.9	18.3	0.4
5	IMes	18.6	20.3	1.7	19.0	18.7	−0.3

---

Finally, to explore the limits of the Mes-substituted NHC copper complexes in N₂O deoxygenation catalysis, 4* and 5* were examined with a reduced catalyst loading of 0.05 mol% (0.05 mM) in combination with B₂pin₂ as the reducing agent and THF as the solvent at room temperature and p_N2O = 1 bar gauge (Scheme 2). Under these mild conditions both pre-catalysts delivered >95% conversion after 20 h in the dark, corresponding to ∼2000 TON and TOF_avg ∼ 100 h⁻¹. This is a step-change in catalytic performance and productivity compared to our previous benchmarks of TON ∼ 850 and TOF_avg ∼ 35 h⁻¹ set using 2* at 80 °C.⁸ Emphasising the importance of excluding light, only ∼5% conversion was observed when these reactions were repeated in a glass pressure reactor exposed to ambient light throughout the experiment.

Scheme 2 High turnover conditions for the catalytic deoxygenation of N₂O.

Conclusions

The catalytic deoxygenation of N₂O mediated by copper(I) complexes stabilised by Dipp- and Mes-substituted NHC ligands has been examined under a variety of reaction conditions, including different diboron(4) reducing agents, and benchmarked against a rhodium(I) system. While boryl derivatives are unstable and decompose rapidly in light, copper(I) pre-catalysts ligated by SIMes and IMes deliver the highest catalytic activity in combination with B₂pin₂ as the reducing agent when performed in the dark using THF as the solvent, achieving ∼2000 TON over 20 h at room temperature under 1 bar gauge of N₂O pressure. DFT-based computational analysis of the copper(I)-boryl catalysed reaction corroborates a mechanism involving reaction with N₂O by O-atom insertion into the Cu–B bond (via initial κ_N-coordinaton) followed by sigma-bond metathesis between the resulting copper(I) boroxide and diboron(4) reducing agent,²⁰ with the relative barriers nuanced by the nature of the supporting ligand and solvent employed. These results further highlight how late transition metal boryl complexes can be used as efficient catalysts for the transformation of N₂O and provide important insights into how the activity of copper(I) boryl systems, in general, can be optimised by careful consideration of the reaction conditions, especially the exclusion of light.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: full experimental details and selected data; computational details and further analysis; optimised geometries in XYZ format. See DOI: https://doi.org/10.1039/d6cy00561f. CCDC 2549087 and 2549088 contain the supplementary crystallographic data for this paper.²¹

Acknowledgements

This work was supported by NERC grants NE/S007350/1 (CENTA2 studentship to ACCW) and NE/X018377/1. We also acknowledge funding from the Leverhulme Trust (RPG-2022-214, TMH) and the University of Warwick. Computational resources and technical support provided by DJEI/DES/SFI/HEA Irish Centre for High-End Computing (ICHEC) and LuxProvide.

References

1. A. R. Ravishankara, J. S. Daniel and R. W. Portmann, Science, 2009, 326, 123–125; Sixth Assessment Report from the IPCC, 2021, https://www.ipcc.ch/report/ar6/wg1/, (retrieved 07/04/2026).
2. W. B. Tolman, Angew. Chem., Int. Ed., 2010, 49, 1018–1024; K. Severin, Chem. Soc. Rev., 2015, 44, 6375–6386; X. Wu, J. Du, Y. Gao, H. Wang, C. Zhang, R. Zhang, H. He, G. Lu and Z. Wu, Chem. Soc. Rev., 2024, 53, 8379–8423; A. Genoux and K. Severin, Chem. Sci., 2024, 15, 13605–13617.
3. G. A. Vaughan, P. B. Rupert and G. L. Hillhouse, J. Am. Chem. Soc., 1987, 109, 5538–5539; P. T. Matsunaga, G. L. Hillhouse and A. L. Rheingold, J. Am. Chem. Soc., 1993, 115, 2075–2077; A. W. Kaplan and R. G. Bergman, Organometallics, 1997, 16, 1106–1108; A. W. Kaplan and R. G. Bergman, Organometallics, 1998, 17, 5072–5085; J.-H. Lee, M. Pink, J. Tomaszewski, H. Fan and K. G. Caulton, J. Am. Chem. Soc., 2007, 129, 8706–8707; S. W. Kohl, L. Weiner, L. Schwartsburd, L. Konstantinovski, L. J. W. Shimon, Y. Ben-David, M. A. Iron and D. Milstein, Science, 2009, 324, 74–77; D. J. Mindiola, L. A. Watson, K. Meyer and G. L. Hillhouse, Organometallics, 2014, 33, 2760–2769; L. E. Doyle, W. E. Piers and J. Borau-Garcia, J. Am. Chem. Soc., 2015, 137, 2187–2190; F. L. Vaillant, A. M. Calbet, S. González-Pelayo, E. J. Reijerse, S. Ni, J. Busch and J. Cornella, Nature, 2022, 604, 677–683; A. Mateos-Calbet, P. C. Bruzzese, M. A. Mermigki, A. Schnegg, D. A. Pantazis and J. Cornella, J. Am. Chem. Soc., 2025, 147, 19438–19443.
4. T. L. Gianetti, S. P. Annen, G. Santiso-Quinones, M. Reiher, M. Driess and H. Grützmacher, Angew. Chem., Int. Ed., 2016, 55, 1854–1858; R. Zeng, M. Feller, Y. Ben-David and D. Milstein, J. Am. Chem. Soc., 2017, 139, 5720–5723; I. Ortega-Lepe, P. Sánchez, L. L. Santos, P. Lara, N. Rendón, J. López-Serrano, V. Salazar-Pereda, E. Álvarez, M. Paneque and A. Suárez, Inorg. Chem., 2022, 61, 18590–18600; S. T. Nappen, J. J. Gamboa-Carballo, E. Tschanen, F. Ricatto, M. D. Wörle, A. Thomas, M. Trincado and H. Grützmacher, Angew. Chem., Int. Ed., 2025, 64, e202502616; S. H. Dewick, T. M. Hood, Y. Han, S. Huband and A. B. Chaplin, Catal. Sci. Technol., 2025, 15, 4126–4129.
5. S. I. Kalläne, T. Braun, M. Teltewskoi, B. Braun, R. Herrmann and R. Laubenstein, Chem. Commun., 2015, 51, 14613–14616.
6. Y. R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, 2007.
7. D. S. Laitar, P. Müller and J. P. Sadighi, J. Am. Chem. Soc., 2005, 127, 17196–17197.
8. T. M. Hood, R. S. C. Charman, D. J. Liptrot and A. B. Chaplin, Angew. Chem., Int. Ed., 2024, 63, e202411692.
9. C. Borner, L. Anders, K. Brandhorst and C. Kleeberg, Organometallics, 2017, 36, 4687–4690; W. Drescher, C. Borner and C. Kleeberg, New J. Chem., 2021, 45, 14957–14964; T. M. H. Downie, R. S. C. Charman, J. W. Hall, M. F. Mahon, J. P. Lowe and D. J. Liptrot, Dalton Trans., 2021, 50, 16336–16342.
10. J. Beaudelot, S. Oger, S. Peruško, T.-A. Phan, T. Teunens, C. Moucheron and G. Evano, Chem. Rev., 2022, 122, 16365–16609.
11. N. P. Mankad, D. S. Laitar and J. P. Sadighi, Organometallics, 2004, 23, 3369–3371; D. S. Laitar, PhD thesis, Massachusetts Institute of Technology, USA, 2006, p. 43; G. G. Dubinina, J. Ogikubo and D. A. Vicic, Organometallics, 2008, 27, 6233–6235; O. Santoro, F. Lazreg, Y. Minenkov, L. Cavallo and C. S. J. Cazin, Dalton Trans., 2015, 44, 18138–18144.
12. M. Teltewskoi, J. A. Panetier, S. A. Macgregor and T. Braun, Angew. Chem., Int. Ed., 2010, 49, 3947–3951.
13. W. Drescher and C. Kleeberg, Inorg. Chem., 2019, 58, 8215–8229.
14. M. A. Esteruelas, M. Oliván and A. Vélez, Organometallics, 2015, 34, 1911–1924; J. J. Gair, Y. Qiu, N. H. Chan, A. S. Filatov and J. C. Lewis, Organometallics, 2017, 36, 4699–4706.
15. S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200–1211; J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33, 8822–8824; C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785–789; A. D. Becke, Phys. Rev. A: At., Mol., Opt. Phys., 1988, 38, 3098–3100; A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652; F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305; F. Weigend, Phys. Chem. Chem. Phys., 2006, 8, 1057–1065; A. V. Marenich, C. J. Cramer and D. G. Truhlar, J. Phys. Chem. B, 2009, 113, 6378–6396; S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104; S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465.
16. H. Zhao, Z. Lin and T. B. Marder, J. Am. Chem. Soc., 2006, 128, 15637–15643.
17. X. Guo and Z. Lin, Chem. Sci., 2024, 15, 3060–3070; X. Guo, T. Yang, F. K. Sheong and Z. Lin, ACS Catal., 2021, 11, 5061–5068.
18. T. Ziegler and A. Rauk, Theor. Chim. Acta, 1977, 46, 1–10; A. Michalak, M. Mitoraj and T. Ziegler, J. Phys. Chem. A, 2008, 112, 1933–1939; M. P. Mitoraj, A. Michalak and T. Ziegler, J. Chem. Theory Comput., 2009, 5, 962–975; L. Zhao, M. von Hopffgarten, D. M. Andrada and G. Frenking, WIREs Comput. Mol. Sci., 2018, 8, e1345.
19. V. Zhuravlev and P. J. Malinowski, Angew. Chem., Int. Ed., 2018, 57, 11697–11700.
20. Our computational analysis concurs with an independent study published during the preparation of this manuscript, F. Shiri and Z. Lin, Inorg. Chem., 2026, 65, 9617–9624.
21. CCDC 2549087: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2rkjmh; CCDC 2549088: Experimental Crystal Structure Determination, 2026,  DOI:10.5517/ccdc.csd.cc2rkjnj.

---